Google launches Veo video model on Vertex AI

Google Cloud announced on Tuesday that Veo, its advanced model for generating text-to-video and image-to-video content, is now available in private preview on Vertex AI. This marks a key addition to Google’s generative AI and machine learning platform.

Originally introduced in May, Veo was developed by DeepMind, Google’s research unit. The model transforms AI-generated images into videos and creates short video clips directly from text prompts.

Starting next week, Google Cloud customers will also gain access to Imagen 3, a powerful text-to-image generator, on Vertex AI. New features for Imagen 3, such as Editing and Customization, will be available to users on the Google allowlist. This curated list ensures that approved users can explore advanced features within Vertex AI.

These updates highlight Google’s commitment to advancing its AI tools. They follow a February incident where the Gemini model’s imaging system created inaccurate depictions of historical figures. Google continues to refine its technologies to provide reliable and cutting-edge AI solutions.

Competition against Adobe

Google has made strides since the Gemini model malfunction, and Veo stands as one of the most advanced imaging technologies. However, Adobe remains the dominant player in the generative AI imaging market, according to Andy Thurai, an analyst at Constellation Research. Adobe’s Firefly tool, widely used by advertisers and creatives, solidifies its market lead.

“Adobe has cornered this market,” Thurai stated. “It will be challenging for Google or OpenAI to break into the creative use case, which Adobe completely owns right now.” He suggested that Google or OpenAI, known for its DALL-E imaging system, might benefit from partnering with a major player rather than competing head-on.

One key factor driving Adobe’s success is its focus on IP safety, noted Keith Kirkpatrick, an analyst at Futurum Group. Adobe guarantees its tools do not infringe on copyright or intellectual property rights and even indemnifies users against legal claims for improper use of original content.

“Both creatives and business managers responsible for content want assurance that generated materials are IP-safe,” Kirkpatrick explained. “No one wants legal issues over generated content.”

In response, Google embeds digital watermarking in every image and video generated by Veo and Imagen 3. The company also plans to indemnify Veo outputs on Vertex AI once the tool becomes generally available. Additionally, Google equips its models with safeguards to prevent harmful content creation.

Despite these measures, Adobe’s reputation for responsible AI remains stronger. The company trains its models using vetted, licensed data, such as Adobe Stock images. “That’s a powerful message to the market,” Kirkpatrick said. “Can Google reach that level? Yes, but it will take time.”

Veo — **Source – cloud.google.com, Image Address – https://storage.googleapis.com/gweb-cloudblog-publish/original_images/1_blog-veo-i2v-1.gif**

Veo’s Advantage

Many creatives are exploring Veo because it offers a key advantage: the ability to produce video clips longer than a minute, compared to Firefly’s five-second limit, according to Keith Kirkpatrick.

Veo’s ability to interpret complex text also sets it apart, noted Andy Thurai. “Most other models struggle to accurately understand and execute based on the user’s intent,” he explained.

To stand out further in the image and video generation market, Google could partner with a major player in the entertainment industry, suggested Michael Bennett, AI policy adviser at Northeastern University’s Institute for Experiential AI.

“I’m waiting for one of the big players to team up with a Hollywood heavyweight,” Bennett said. “A major announcement like that would signal a real tectonic shift in the field.”